Data collection methods

Susan Sullivan-Bolyai and Carol Bova

Learning outcomes

After reading this chapter, you should be able to do the following:

• Define the types of data collection methods used in nursing research.

• List the advantages and disadvantages of each data collection method.

• Compare how specific data collection methods contribute to the strength of evidence in a research study.

• Identify potential sources of bias related to data collection.

• Discuss the importance of intervention fidelity in data collection.

• Critically evaluate the data collection methods used in published research studies.

Go to Evolve at http://evolve.elsevier.com/LoBiondo/ for review questions, critiquing exercises, and additional research articles for practice in reviewing and critiquing.

Nurses are always collecting information (or data) from patients. We collect data on blood pressure, age, weight, height, and laboratory values as part of our daily work. Data collected for practice purposes and for research have several key differences. Data collection procedures in research must be objective and systematic. By objective, we mean that the data collected are free from the researchers’ personal biases, beliefs, values, or attitudes. By systematic, we mean that the data are collected in a uniform, consistent, or standard way from each subject by everyone who is involved in the data collection process. When reading a study, the data collection methods should be identifiable and repeatable. Thus, when reading the research literature to inform your evidence-based practice, there are several issues to consider regarding study data collection descriptions.

It is important that researchers carefully define the concepts or variables they are interested in measuring. The process of translating a concept into a measurable variable for data collection requires the development of an operational definition. An operational definition is how the researcher will measure each variable. For example, Thomas and colleagues (2012) (see Appendix A) defined quality of life as including the four domains of physical, social, emotional, and functional well-being measured by the FACT-G, which is commonly used in cancer studies.

Ultimately, the degree to which the researcher is able to clearly define and measure study variables in an unbiased and consistent way is an important determinant of how useful the data will be to guide practice. The purpose of this chapter is to familiarize you with the various ways that researchers collect data from subjects. The chapter provides you with the tools for evaluating the types of data collection procedures commonly used in research publications, their strengths and weaknesses, how consistent data collection operations (a high level of fidelity) can increase study rigor and decrease bias that affects study internal and external validity, and how useful each technique is for providing evidence for nursing practice. This information will help you critique the research literature and decide whether the findings provide evidence that is applicable to your practice setting.

Measuring variables of interest

To a large extent, the success of a study depends on the fidelity (consistency and quality) of the data collection methods used. Researchers have many types of methods available for collecting information from subjects. Determining what measurement to use in a particular study may be the most difficult and time-consuming step in study design. Thus the process of evaluating and selecting the available instruments to measure variables of interest is of critical importance to the potential success of the study.

As you read research articles and the data collection techniques used, look for consistency with the study’s aim, hypotheses, setting, and population. Data collection methods may be viewed as a two-step process. First, the researcher chooses the data collection method(s) for their study. In this chapter the selection of measures and the implementation of the data collection process are discussed. An algorithm that influences a researcher’s choice of data collection methods is diagrammed in the Critical Thinking Decision Path. The second step is deciding if these methods are reliable and valid. Reliability and validity of instruments are discussed in detail in Chapter 15 (for quantitative research) and in Chapter 6 (for qualitative research).

CRITICAL THINKING DECISION PATH

Consumer of Research Literature Review

Data collection methods

When reading a research article, be aware that investigators must decide early in the research process whether they need to collect their own data or whether data already exist in the form of records or databases. This decision is based on a thorough literature review and the availability of existing data. If the researcher determines that no data exist, new data can be collected through observation, self-report (interviewing or questionnaires), or by collecting physiological data using standardized instruments or testing procedures (e.g., laboratory tests, x-rays). Existing data can be collected for research purposes by extracting data from medical records or state and national databases using standardized procedures. Each of these methods has a specific purpose, as well as pros and cons inherent in its use. It is also important to remember that all data collection methods rely on the ability of the researcher to standardize these procedures to increase data accuracy and reduce measurement error.

Measurement error is the difference between what really exists and what is measured in a given study. Every study has some amount of measurement error. Measurement error can be random or systematic (see Chapter 15). Random error occurs when scores vary in a random way. Random error occurs when data collectors do not use standard procedures to collect data consistently among all subjects in a study. Systematic error occurs when scores are incorrect but in the same direction. An example of systematic error occurs when all subjects were weighed using a weight scale that is under by 3 pounds for all subjects in the study. Researchers attempt to design data collection methods that will be consistently applied across all subjects and time points to reduce measurement error.

Fidelity means that data are collected from each subject in exactly the same manner with the same method by carefully trained data collectors (see Chapter 8). To help you decipher the quality of the data collection section in a research article, we will discuss the three main methods used for collecting data: observation, self-report, and physiological measurement.

HELPFUL HINT

Remember that the researcher may not always present complete information about the way the data were collected, especially when established instruments were used. To learn about the instrument that was used in greater detail, you may need to consult the original article describing the instrument.

EVIDENCE-BASED PRACTICE TIP

It is difficult to place confidence in a study’s findings if the data collection methods are not consistent.

Observational methods

Observation is a method for collecting data on how people behave under certain conditions. Observation can take place in a natural setting (e.g., in the home, in the community, on a nursing unit) or laboratory setting and often includes collecting data on communication (verbal, nonverbal), behavior, and environmental conditions. Observation is also useful for collecting data that may have cultural or contextual influences. For example, if a nurse researcher wanted to understand the emergence of obesity among immigrants to the United States, it might be useful to observe food preparation, exercise patterns, and shopping practices among specific immigrant groups.

Although observing the environment is a normal part of living, scientific observation places a great deal of emphasis on the objective and systematic nature of the observation. The researcher is not merely looking at what is happening, but rather is watching with a trained eye for specific events. To be scientific, observations must fulfill the following four conditions:

1. Observations undertaken are consistent with the study’s aims/objectives.

2. There is a standardized and systematic plan for observation and data recording.

3. All observations are checked and controlled.

4. The observations are related to scientific concepts and theories.

Observational methods may be structured or unstructured. Unstructured observation methods are not characterized by a total absence of structure, but usually involve collecting descriptive information about the topic of interest. In participant observation, the observer keeps field notes (a short summary of observations) to record the activities, as well as the observer’s interpretations of these activities. Field notes usually are not restricted to any particular type of action or behavior; rather, they represent a narrative set of written notes intended to paint a picture of a social situation in a more general sense. Another type of unstructured observation is the use of anecdotes. Anecdotes are summaries of a particular observation that usually focus on the behaviors of interest and frequently add to the richness of research reports by illustrating a particular point (see Chapters 5 and 6 for more on data collection strategies). Structured observations involve specifying in advance what behaviors or events are to be observed. Typically, standardized forms are used for record keeping and include categorization systems, checklists, or rating scales. Structured observation relies heavily on the formal training and standardization of the observers (see Chapter 15 for an explanation of interpreter reliability).

Observational methods can also be distinguished by the role of the observer. The observer’s role is determined by the amount of interaction between the observer and those being observed. These methods are illustrated in Figure 14-1. Concealment refers to whether the subjects know they are being observed, and intervention deals with whether the observer provokes actions from those who are being observed. Box 14-1 describes the four basic types of observational roles implemented by the observer(s). These are distinguishable by the amount of concealment or intervention implemented by the observer.

BOX 14-1 BASIC TYPES OF OBSERVATIONAL ROLES